| name | capacity |
| description | Show local container engine capacity and HAL deployment estimates. Use when the user asks about available resources, whether there is headroom for a new stack, or what is currently deployed and how heavy it is. |
HAL Capacity Skill
The hal capacity command surfaces live engine resource usage and per-stack footprint estimates so you can judge headroom before spinning up a new product stack.
Primary Commands
hal capacity # current engine view: CPU, RAM, running containers
hal capacity --active # active heavy deployment composition with per-stack footprint details
hal capacity --pending # pending heavy deployment impact estimates
When To Use
- Before
hal terraform create or any other heavy create — to check if the engine has enough headroom.
- After noticing sluggishness — to see which stacks are consuming the most resources.
- To understand relative cost of stacks before deciding what to tear down.
Key Behaviours
- Memory pressure calculations exclude cache/buffers (pressure memory, not free-cache-inflated baselines).
- Interactive confirmation prompts on heavy creates only trigger when estimated post-create usage exceeds engine limits (CPU > 100% or RAM > machine RAM).
- Podman on macOS exposes richer machine-runtime telemetry than Docker when available.
- Capacity scenario labels are infra-centric (e.g. shared GitLab runner, KinD/VSO flows), not purely product-centric.
Lab Assumptions
hal status also surfaces a compact capacity summary alongside product state.
- Use
hal capacity --active to see exactly which running stacks own the current resource footprint before deciding what to scale down.